One big problem I have with these decline numbers is that the returns on in field drilling are expected to be constant. And in field drilling itself is expected to occur at a constant rate.

It makes more sense to assume in field drilling is influenced by price and the field decline.
Higher prices accelerate in field drilling and declines accelerate in field drilling.

Also in field drilling by definition is finite and can only expand until the field is fully drilled.
Your ability to expand and maintain your infield drilling campaign is limited to the size of the
field.

Given the above we would expect that over the last several years in field drilling campaigns have been steeped up and at some point will result in steeper decline rates as they have increased the depletion
rate.

You saw a similar pattern when the US peaked except without the technical advances that keep production higher to greater depletion levels.

Whats really needed to understand our future oil supply given that discovery is well in the past is a understanding of the infield drilling campaigns and their effect at the field level.

Given that most infield drilling campaigns implicitly work to keep production at its peak design level and generally don't exceed it by to much. One would expect that producers of existing fields will do whatever they can to keep production close to peak but the pressure to increase production in existing fields is low simply because of constraints on the above ground oil gathering infrastructure. You can have production decline and deal with that fairly easily but expanding production is exponentially more expensive then maintaining the current production rate.

The overall effect is you have an unknown change in the depletion rate driving a constant production rate.

However one thing is for sure if production remains constant then the depletion rate is increasing in developed fields.

Next we know for a fact that our technology is capable of extracting oil at high depletion rates 20-25% is not unknown and in some fields even higher. Thus our ability to deplete a oil field with modern technology is probably close to physical limits increasing depletion rates beyond 20% or so becomes limited by EROEI issues.

This can be seen in the steady decline in field lifetimes over the last few decades generally blamed on finding smaller fields but we know from WHT's work that discovery does not follow field size so this is a incorrect assumption. Instead given everything we know we should expect that field depletion rates have been climbing on average for decades. And further more we know that the maximum depletion rates possible are high.

Now the way around this situation is to increase URR and thus decrease the calculated depletion rate
the easiest way to increase URR in existing fields is to type a new entry into a computer database.

Problem solved.

Memmel, this is similar to what I thought: The IEA numbers only distinguish natural decline and observed decline. But for a forecast (or scenario) both are only theoretical numbers, which provide an orientation:
Natural decline only happens if there is zero additional investment. But in reality this rarely happens as long as the field isn't hopelessly depleted - except for serious above-ground problems like in Iraq.
Also the "observed decline" only gives an orientation from the historical development, as the future decline doesn't simply depend on *if* investments will be made but also *which* and *how many* investments are made. For example future decline may depend on if the remaining reserve of a region will be tackled by vertical or horizontal wells. Much of this will depend on the future oil price, which may determine if more expensive methods will pay off - or also if sufficient capital, equipment, personnel etc. are available. In their scenarios this is partly addressed by the the IEA as they distinguish between conventional crude and "EOR" reserves.
So I don't think that the IEA's decline numbers can only be used as a rough orientation but not as a forecast as future production depends on many more parameters.